position-dodge

Dodge position adjustment: place grouped marks side by side.

Typically set on a layer as position: "dodge" rather than constructed directly; the constructor exists for symmetry with the other positions. When all marks at a given x share the same width, the result matches a simple uniform dodge. When widths differ (per-row width column), each mark uses its own width as its slot, with padding between slots and a shrink-to-fit if total slot use would exceed the bucket.

Usage

position-dodge(
  width: 0.9,
  padding: 0.1,
)

Parameters

Parameter	Default	Description
width	0.9	Total width reserved for the dodged group, as a fraction of the category width.
padding	0.1	Gap between adjacent dodge slots in mixed-width mode, as a fraction of the bucket.

Returns

Position dictionary with name: "dodge", consumed by plot.

Examples

Bars dodged side by side per fill group within each x slot.

#let d = (
  (q: "Q1", grp: "a", y: 3),
  (q: "Q1", grp: "b", y: 5),
  (q: "Q2", grp: "a", y: 4),
  (q: "Q2", grp: "b", y: 2),
)
#plot(
  data: d,
  mapping: aes(x: "q", y: "y", fill: "grp"),
  layers: (geom-col(position: "dodge"),),
  width: 10cm,
  height: 6cm,
)

Increase padding to widen the gap between dodged bars.

#let d = (
  (q: "Q1", grp: "a", y: 3),
  (q: "Q1", grp: "b", y: 5),
  (q: "Q2", grp: "a", y: 4),
  (q: "Q2", grp: "b", y: 2),
)
#plot(
  data: d,
  mapping: aes(x: "q", y: "y", fill: "grp"),
  layers: (geom-col(position: position-dodge(padding: 0.3)),),
  width: 10cm,
  height: 6cm,
)

See also

position-stack, position-fill, position-identity, position-jitter.